Chinese AI chipmaker Transtreams (凌川科技) has raised approximately $27.5 million in a series A funding round. The funding was co-led by the Beijing Artificial Intelligence Industry Investment Fund and Kuaishou, with contributions from Yizhuang Industrial Investment, Shunxi Fund, and Jiuzhi Capital.

Transtreams, founded in March 2024 as a spin-off from Kuaishou, plans to use the funds to develop next-generation chips and scale production of its SL200 chip.

Additionally, the company aims to expand into international markets.

The SL200 chip is currently deployed in data centers and various industries. It integrates video encoding, AI inference, and multi-core CPU functionalities. Its applications include smart cities, intelligent inspections, and support for firms such as Kuaishou, Alibaba, and Baidu. It is also being utilized in Southeast Asia and Brazil through Kuaishou’s operations.

The company has completed compatibility tests for a new chip designed for large language model training and inference. Lingchuan Tech aims to expand into autonomous driving and edge computing in the near future.

🔗 Source: 36Kr

🧠 Food for thought

1️⃣ China’s strategic push into application-specific AI chips counters Nvidia’s dominance

Lingchuan Tech represents China’s emerging strategy to compete in the AI chip market by focusing on specialized applications rather than directly challenging Nvidia’s general-purpose dominance.

While Nvidia controls 70-95% of the overall AI chip market with its powerful but expensive GPUs costing upwards of $30,000 each, Chinese startups are targeting specific use cases like video processing where specialized designs can offer cost advantages 1.

This application-specific approach allows newer entrants to carve out market share in the projected $400 billion AI chip market without directly competing with Nvidia’s strengths in training large language models 2.

Lingchuan’s emphasis on cost efficiency for video tasks aligns with broader industry recognition that the transition from AI training to inference presents opportunities for alternatives to Nvidia’s expensive solutions 1.

The company’s Chinese government backing through the Beijing Artificial Intelligence Industry Investment Fund follows a pattern of strategic national investment to develop domestic chip capabilities, similar to government funding that historically accelerated computing research in the United States 3.

2️⃣ Video processing emerges as strategic entry point for AI chip startups

Lingchuan Tech’s focus on intelligent video processing highlights how computational video has become a critical entry point for new AI chip companies seeking established use cases with immediate commercial applications.

The company’s SL200 chip, already handling video tasks for 700 million users through Kuaishou’s platform, demonstrates the massive scale and immediate commercial demand for video-specific AI processing that new chip architectures can address 4.